LISTing Newsletter

Issue! DECEMBER 1996 LISTing Policy ........ Annual Dues $ 16.69

On e sample copy sent upon receipt of business size SASE. Copies provided on

Exchange basis with. other Bona fide ahi We are always lalina be articles,

pro: oe etc to keep members informed and entertained. You maintain full credit

co,

P hortos of this J oat ai may be reproduced, without need for written consent
in,

Please give credit to LISTing when reprinting any articles. List disclaims responsibility for
any damage to your computer as a result of. polit any articles in LISTing.

PERSONAL COMPUTERS

Computing Computing’s True Cost

Corporate computer managers have known it for years, but most
individual personal computer buyers learn it the hard way: the purchase
of one PC is just the tip of the iceberg.

Many families are saving to buy a new computer for the holidays. But
the excitement of getting the new plaything can evaporate quickly if
one does not account for the myriad additional expenses that it brings.
For those who will be receiving a PC from Santa, remember: a computer
is a gift that keeps on taking.

Beyond the obvious extra outlay for sales tax, other hidden or
camouflaged costs can surprise, disappoint and ultimately dismay the
buyer. In general, if a new personal computer costs $2,000.00, it
will demand another $2,000.00 worth of software, peripherals and toys.

The budget creep begins on the first day of shopping: First, the
advertisement may show a computer with a CD-ROM drive, a big screen
monitor, a keyboard and mouse and perhaps a printer, topped with a

sign that reads, "Only $1,495.00" or something like that. Grab the
magnifying glass. The small print typically advises, "monitor and
printer not included". Monitors and printers add at least a couple of

hundred dollars each to the system price.

We tend to think poorly of used-car salesmen, but at least they do not
ambush us with "Steering wheel and tires extra". Or the macho-looking
PC is shown with two headlines, one that reads, "250 Megahertz!" and
the other, "Starting at $1,999.00!" Do not assume the two selling
points are related. Again, the fine print reveals that the $1,999.00
model uses the runt littermate of the fast chip, or has only half the
memory of the bigger model, or has the bargain-bin 4X CD-ROM drive, and
so on.

By the time the sales person persuades you that you really want the
better components, the extra memory, the modem, the year or two of
onsite service, and the surge protector, and the box of diskettes, and
the external speakers, and ~ what the heck - the mousepad, and the
wrist rest, and the joystick, the video eyeball gizmo, and the special
printer paper, the scanner, the special ergonomic chair and computer
desk, the propeller beanie ... Yow! Your budget is in shreds.

Most of these things are subject to my 18@-day rule: As soon as you
determine you absolutely have to have something, put it on a list and
date it. Wait six months. By then, you will have discovered either
that you really don’t need it or that its price has come down and the
bugs have been fixed.

Some extra dollars, however, are well spent at the time the computer is
bought. Although many new PCs come with 16 megabytes of system memory,
there are a few that still come with 8. Meanwhile, the software
companies continue to crank out bloatware, and there are no signs the
trend to memory-hogging software is abating.

Fortunately, memory chip prices are very attractive right now. One can
add chunks of memory in 8 megabyte increments for about $50.00 each.
Consider equipping the new PC with 32 megabytes of RAM while the price
is right. (continued on page 10)

PAGE - 2 -

Qi CORNER

Last month I stated that I had purchased a clear Black light kit for
use aS an eprom eraser. The kit was assembled and did NOT do the job
of erasing eproms. I gave the kit to an eleven year old boy who lives
next door to me. He use will be for locating invisible flurescent
minerals. I hope he has better luck with it than I had with it!

I use my QEP eprom programmer quite often, and occasionally I have an
error in the program within the eprom. Out with my old eprom eraser
which takes approximately twenty-five minutes to erase a 27C512 eprom.
This is a long time to erase the entire chip.

I have a pretty good library of articles for all sorts of electronic

items. Searching back for some articles which were written
approximately 15 years, I stumbled upon a letter from a friend of mine
from Canada. He had experimented with various lamps for erasing

eproms. Here it is in it’s entirety, a simple and inexpensive way to
erase eproms. The author does not want to be acknowledged!

Virtually all eproms require a light source of 2537 Angstroms to erase
the contents back to its original state of all outputs high or all
outputs low. For a long time the only method of getting that type of
light source that I new of was to buy a commercial eprom eraser. By
the time I found a cheap eprom eraser, I was looking at over one
hundred dollars. Since then I’ve found you can have an eraser for as
low as $27.00 (this was January 1981).

At 2537 Angstroms light does not pass through common glass or the
atmosphere very well. This means ordinary mercury vapor ultra-violet
sources such as the ones you use for exposing sensitized PC boards or
making blueprints do not work when it comes to erasing eproms.
Exposing to direct sunlight doesn’t seem to work either because of the
atmospheric absorption of light rays in this region. I can safely say
this because I’ve tried these methods including "black lights" and
Plant lights.

After searching around I finally found the G8T5 bulb (better known as
germicidal lamp). The beautiful part of this lamp is that it will fit
any standard desk top flourescent lamp that uses a 12" bulb. All you
have to do is replace the existing bulb with the G8T5 or equivalent.
The bulb isn’t cheap so you had better not drop it too many times. The
secret to this bulb is quartz tubing instead of glass.

Placed at 3/4" away from the eprom window it will erase the eprom
contents in about 18 minutes and even less for a 2716 eproms. Because
of varying situations, i.e., age deterioration, dust particles on the
eprom window, centering of the light source, etc., you should runa
complete "read" of all the addresses on the newly erased eprom to
ensure there is not a partial erase on any of the bits. It can be very
frustrating to find a partially erased bit at a location after you’ve
keyed in everything correctly up to that point.

Warning! Although light sources in this range of the spectrum may not
look very dangerous or even very bright, it can cause skin cancer under
continued exposure. At 2537A, it is well beyond the light spectrum you
can see but it will still destroy the cells in your eye, eventually
causing blindness. The moral of this story is to clearly label a
warning on your lamp fixture and don’t try to read with this bulb!

PAGE - 3 -

For safety reasons or if you don’t have a suitable lamp fixture already
you can build an inexpensive unit such as shown in the drawing below.
It’s simply a piece of .032 aluminum folded into an inverted ’U’ trough
with 3/4" hardwood plywood ends mitered as shown to keep the light from
escaping at the ends. The wiring is shown in figure A. Simply center
this light source over your eproms, turn it on and come back after the
required exposure time. If you’re building from scratch, be sure to
mount the lamp sockets so that there is a least 3/4" clearance between
the bottom of the lamp and the base line. You can even save on sockets
by soldering the leads directly to the bulb.

ON

GE-89G6435 ofr

The Eprom

The first memory devices for computers were bi-stable vacuum tube
circuits known as ’filp flops’ and sonic pulses circulating through a
length of metal or a tube of mercury. The first memory device that
stayed around awhile was the "core" memory. This used a little donut
of ferrite that could be magnetically polarized by a pulse of current
in a wire passing through its center. The problem with this was that
you could only tell which way it was polarized by giving it another
pulse big enough to change its polarization and thus erasing the stored

data.

These little donuts (cores) were woven into “mats” with three wires
through each core, two for select and one to read the condition.
Whenever you read out a bit, you had to rewrite it if you wanted it to
be there. In spite of this complexity, core memories became the
dominant form till the semiconductor memory replaced it in the early
7@’s. The first semiconductor memories were like the vacuum type. A
bunch of flip flops, but made very small.

In rapid sequence Intel introduced the first SRAM (Static Random Access
Memory) and the DRAM accesses the same as the SRAM, but has a simpler
memory element. The DRAM accesses the same as the SRAM but has a
simpler memory element. It operates by storing the data as a charge on
a tiny capacitor. This allowed a smaller memory cell and thus a larger
memory matrix for a given chip size, but because of the small size of
the storage capacitor, the memory had to be "refreshed" several hundred
times a second. That is; whatever state it was in, it had to be
rewritten to enhance that state. At first this seemed to many to be
a precarious way to store data, but the DRAM proved reliable and is now
the most common memory element.

PAGE - 4 -

In 1972, very shortly after the introduction of the DRAM, Intel also
developed the EPROM (Erasable Programmable Read Only Memory). This
device uses a single MOS transistor as the memory element. It is
equipped with a "floating gate" that controls whether it is conducting
or not. This gate is like a little capacitor plate over the transistor
that is completely surrounded by insulating oxide. Because there is
normally no conductive path to it, a charge or voltage level will
remain indefinitely. The trick is to have a connection for normal use.
This is done by making the insulating oxide thin enough, and uniform
enough in its thinness, that a higher than normal voltage applied will
cause a nondestructive break down or leak that will change the voltage
level of the floating gate. If it can be changed both ways it is
called an EEPROM (Electrically Erasable PROM) but this proved very
difficult and only recently EEPROMs (or E2PROM) become competitive.
The single direction gate charge was easier and another way was found
to reverse the effect; UV light.

The charge that could be put on the floating gate could be removed by
exposing the transistor to 254 nm wavelength UV light. There is
nothing magic about. 254mm. It just happens to be the easiest wave
length to produce with a mercury vapor lamp and conveniently is nearly
absent from sunlight and common light sources.

A 27256 EPROM has 262,144 bits of memory like the 256K DRAM, but since
it is organized as Byte wide data, there are only 256K/8 = 32K memory
locations and only fifteen address lines are needed. With fifteen
address lines and 8 data lines at least 23 pins are needed. Add power,
and you have 25 pins. There are several other functions needed and
these are cleverly combined in the remaining three pins of a 28 pin
EPROM package.

One of these functions is to allow the data lines to be used as both in
and out. Another is to provide the higher programming voltage needed
for writing to the EPROM. Since all memory devices in a system are
tied to the same address and data lines, control functions are needed
to allow the address information in (usually called "Chip Enable” or
CE) and a means to make the data pins connect or not to the computer
data lines (usually called "Output Enable or OE).

Originally the small capacity EPROMs had all these control functions on
different pins. As more memory was provides on a chip and more address
lines were needed, some of these control pins were required to serve
multiple purposes. ,

Thus, in most EPROMS used now, the control functions are a complicated
combination of voltage levels on these control pins. The 27256 for
example has eight different operating modes all determined by three
control pins.

The EPROM started out ina 24 pin package, as originated by Intel.
Some other manufacturers made EPROMs in packages of fewer pins, but
they have mostly vanished from use and the Intel standard is now almost
universally used.

The first EPROM, the 2708, was a 24 pin device and, as mentioned, was a
straight forward control method. The 2716 was the next and the added
address line meant complicating the control a little. The 2732 was
still a 24 pin package and called for still more combined control pin
action. With the 2764, the package size increased to 28 pins and
simple control again. The 27128 had almost the same control functions

PAGE - 5 -

but with the 27256 it got complicated again. Each of these has a
little different control method for programming, though the "Read"
function is straight forward. Along with this development of larger
size memory matrix, methods were found to reduce the programming
voltage. It went to 25V on the early units to 21V and now to 12.5V.

As stated earlier, an EPROM is read at full computer speed much as any
other memory device. Writing is quite a different situation. The
EPROM data can only be changed from "1’s"“ to "@’s" by writing. "1"
being a high output data line and "®" being a low. High is generally
about 4V and low is less than .5V.

Writing an EPROM is slow compared to computer speeds. It requires
applying programming voltage and data to the EPROM while the desired
address is selected for as much as 50 milliseconds per Byte.

Actually, the way the EPROM’s are written is to apply these signals for
about one millisecond. Then the mode is switched to "Verify" which
allows read out without changing the programming voltage. If the data

read out is not the same as that being written, the-ims write cycle is ~

repeated. This is continued till a proper verify is observed. At that
point the EPROM writing program applies the programming data and
address, as before, but for a period of time that is some function of
the total time it took to get that Byte to absorb the data. . For
example, Intel suggests the final pulse should be three times as long
as the total of all the pulses it took to get a good "Verify" response.
This is to change the floating gate well beyond a threshold condition.

The newer EPROMs have the capability of much faster writing. One
method is to apply 6V to the Vcc pin (normally 5V) and use 100
microsecond programming pulses. This takes advantage of the fact that
most Bytes write much faster than the worst case. Also, operating at
Voc = 6V during programming has something of the same effect as the
final “over program" pulse of the other method. When the "Verify" is
threshold at Vcc = 6V it is well over threshold at Vcc = 5V.

AS mentioned, the EPROM bits can only be written from "1" to "@".
Thus, to write an EPROM it is generally necessary to restore the entire
memory matrix to "1’s" by ultra violet light erasure. If read out in
Hex code, the Bytes appear as FF indicating all bits are "1’s".

EPROM programmers. have the means.of applying the address and the data
codes to the EPROM and holding them there for the relatively long time
necessary. They also have the means of applying the proper program
voltage and control codes to suit the particular EPROM being used.

Normally a programmer provides:

Erase check: To check for full erasure.

Verify: To check the stored data.

Write: To store a block of data.

The Write" function incorporates the first two also. That is; it will
do an Erase check just prior to writing and abort if the EPROM is not
blank. However, most programmers allow you to bypass this routine if
desired so that just a portion of the EPROM can be written.

Ordinarily, on completion of the "Write" cycle, the program does a
"Verify" routine just to be sure it does read out properly. The Erase

PAGE - 6 -

Check and Verify routines can take less than 10 milliseconds for a 256K

EPROM so they are not noticeable

the "Write" code.

Many thanks for the above information from Intel.

See you next month...

A low-cost EPROMer to put your

2X81 to use

ZX-EPROMer

EPROMS are powerful, inexpensive and
familiar components that are found in
many circuits popular today. A
stand-alone programmer can cost over
$1000. A card for your fruit is cheaper.
but then you must have a machine to put
that card in. Stop using your ZX-8] as a
doorstop and cash in on its hidden poten-
tial; build this ‘‘full-feature’’ EPROMer
and enjoy freedom over EPROMs. The
circuit described here will program the
popular 2764 type EPROMs; these are
under $10 and store 8K, making them a
good choice for many applications. Some
uses of this device could be:

1. Program any compatible EPROM.(i.e.,
2716).

2. Read any ROM, PROM, or EPROM.
3. Static RAM in any unoccupied memory
block.

4. A games cartridge system for the
ZX-81.

5. An EPROM copier.

How It Works

While we will be looking at programming
the 2764, the theory of operation is much
the same for other types. It is well worth
vour while to consult the proper manual
before trying a different memory chip.
The programming voltages, for example,
vary from type to type. When programm-
ing the 2764 it needs:

1. Stable data and address lines for the
duration of the write.

2. A stable 21 voltsat 50mA at the Vpp
pin.

3. The PGM and CS held low for 50mS
after 1 and 2 are met.

The OE pin can be tied to CS to allow
for verification of each write. Vpp can be
either 5 or 21 volts while reading; PGM
should be held high when reading. In this
circuit we are slipping our 8K EPROM in-
to the memory map in the ‘‘never

Bob Gilder

ee es

By Nick Dennis

bothered by BASIC’’ 8-16K block. As
such, PEEK and POKE can give us direct
access to all locations. This means that all
software can be written in slow but friend-
ly BASIC.

IC 1 is a block decoder. It looks at
the top three address lines and decodes all
memory requests into 8K wide Block
Selects (8 outputs x 8K ‘ 64K). The YO BS
de-echoes the ROM (familiar ZX stuff)
while its neighbour Y1 delivers the BS
where our EPROM lives.

IC 2 is the trap circuit. If BS is low
and RD is high (this can only mean a write
to the EPROM) and if the timer output is
high, then a WAIT is issued to the Z80.
This has the unadvertised effect of freez-
ing the Z80’s lines until released. The
WAIT is also inverted and after a slight
delay (C3) passed to the first timer’s + ve
trigger. Each timer has two outputs, Qbar
and + Q. Because the 2764 is programmed
with-2 high-low-high sequence PGM is
connected to Qbar. The Q output is ap-

2764
PIN CONFIGURATION

Fig. 1 The pinouts of the 2764 EPROM.

PAGE - 7 -

and David Turner

relative to the minutes necessary for

plied to the second timer’s -ve trigger.
After 50 milliseconds these outputs both
flip, thus ending the programming and
triggering the second timer. The second
timer prevents the LS10 from retrapping
this same write.. But what about
REFRESHing the dynamic RAMs? Even
though D-RAM is usually REFRESHed
every few milliseconds, experience has
shown that good data integrity is main-
tained with occasional periods much
longer than 50mS.

Hooking’er Up
This project is fairly straightforward and
does not require a printed circuit board.
Veroboard is a good choice for most pro-
jects like this. Connecting the veroboard
to a ZX-connect with coloured ribbon
cable will keep wiring simple. As a bonus,
not every edge-connect line has to be
brought out to the board.
A.ZX-connect can be made by cut-
ting a 50 pin 1/10 inch standard edge-con-
nect to length. The key-way tab is made
from veroboard. The expansion strip can
also be made from two pieces of
veroboard sanded to half thickness and
glued back to-back. An easier approach is
to cut the strip from an existing circuit
board. Our strip came from Zebra systems
but we have seen many ideal bits of
fibreglass in our favorite surplus store.
The 1/10 inch spacing is very common.
All this is necessary because the
EPROMer must be the first device plugg-
ed into the expansion port. The Memory
Packs alter the high order address lines (to
be stackable) and this definitely won’t do.
Many peripherals are memory mapped in
the 8-16 K range and as such are not com-
patible with this gadget. To be able to use
a 64K pack with the EPROMer you must
be able to de-select the 8-16 K range.
Most have this feature. Any 1/O mapped

device (e.g., the ZX printer) is cool.

If the expansion strip is not possible
for some reason then bring any altered ad-
dress line (typically A14) directly from in-
side the computer to IC1. This is low-tech
and not too neat, but will still do the trick.

Personality The use of several IC
sockets as ‘‘personality’’ strips is recom-
mended but not vital. The idea here is to
be able to make modifications to control
hookups without soldering. A look at the
pinouts for most popular ICs reveals that
many lines always connect to the same
pins while others change from type to
type. Having the top few address lines and
most of the control lines jumpered
through a row of IC sockets will allow a
lot of scope for testing, adjustment and
hacking. a

The Data iines aiid AO to Aid are
wired directly to the socket. Solder the
control lines as the circuit comes together.
WAIT, RD, MREQ go directly from the
circuit to the ZX-connect. A good ‘‘per-
sonality’’ candidate is WR . While not us-
ed in the circuit, it allows the use of static
RAMs in the socket.

It is worth noting that no points are
awarded for a cramped, tiny board. While
keeping all wire lengths as short as possi-
ble is always a good idea, never paint
yourself into a corner when building
around a flexible circuit like this.

Testing and Adjustment

When your board is all assembled and
well-checked over for solder bridges, it is
time to power it up. Connect it directly to
the ZX edge-connect and (with no
EPROM) apply the power. If the cursor
does not appear re-recheck all wiring.

If it does, try a POKE to anywhere in
the block, e.g., POKE 10000,0. A screen
flicker at this time tells you that all is well.
This is the Z80 ‘‘freezing’’.

If you are unabie to get a screen-

flicker, test the trap by momentarily
grounding 3, 4, and 5 of the LS 10
(disconnect them from LS 138 first). A
screen flicker now means that the problem
is in the block decoding. If you are still
having problems, read the section ‘‘Got-
cha’’.

But you were careful and you got the
screen flicker right away. Centre RI and
R2 and enter the program, except for line
120. The idea here is simple: while 50
milliseconds is hard to time with a stop-
watch, 50 seconds is pretty easy. Discon-
nect the LSI38 from the LSI and (no
flicker) RUN the program. This is approx-
imately 70 seconds in SLOW mode.
Reconnect the LSI38 to the LSIO and
re-RUN the program. For 1000 loops the
‘flicker’ should add 50 seconds to the
“no flicker’? RUN time, giving a total
time of 120 seconds. Adjust RI until this is

Across every IC +-

Fig. 2 The complete circuit of the ZX EPROMer

Pres. . Harvey Rait
Vice Pres. Bob Gilder
Tresurer Robert Malloy
Cor. Secy. John Pazmino
Assoc. Editors Fred Stern
Harvey Rait
Publisher Bob Gilder
Libr. Za a Tom. Skapinski

Please send all inquiries and

submissions (including dues)
to: L.I.S.T. ;
Mr. Harvey Rait
5 Peri Lane,
Valley Stream, N. Y. 11581

COMING EVENTS: The next L.I.S.T.
meeting will be Sunday, 91/12/97
at 2 P.M. at the home of Harvey
Rait (see address above).

PAGE - 8 -

»

2ZX-EPROMer

Nn

exactly right. Now put a drop of your
favourite nail polish on RI’s arm. By now
the use of the ‘‘personality’’ strip should
be apparent. Without it this setup stage
can be a hassle.

R2 is not too critical. If the system
“hangs” on the first write, then increase
its value. Add line 120, delete line 30,
CLEAR the variable area and SAVE the
program to tape. DEST and the number
of LOOPs will change for different ap-
plications, but the theory of operation
and the timing set by RI will remain the

same.

Programming
To simplify programming we will be using
BASIC. Because the socket is

**memory-mapped"’", PEEK and POKE
will give us access to the EPROM. PEEK
behaves normally, while POKE is extend-
ed by the circuit to meet the 50 millisecond
requirement. Remember, POEs are per-
manent! To erase any one location you
must erase the entire EPROM.

The EPROM can be programmed in
many ways or in any order. Most
assemblers have a Block Transfer or
MIOVE function, but usually without
verification. The easy solution is to move
the object code into a string variable and
then type in the minimum BASIC pro-
gram. With I6K of memory there is still
ample room for'most assemblers (4K-7K),
a few lines of BASIC, and 8K of dataina

variable.

To copy an existing EPROM, for ex-
ample, transfer its contents to an array,
save everything to tape, power down, in-
stall the blank, reload and then program.

The BASIC given here is only to
show how simple it is to program an
EPROM. Feel free to include Menus,
features, bells and whistles if you like.
Your program should:

1. Verify that all locations to be pro-
grammed are 255 ($FF) before writing.
You can make a one into a zero but not
vice versa. `.

2. Verify all start/finish values with the
user.

3. Verify that the correct value has been
written. If an error occurs it is not likely
that the machine can correct itself.

Vpp drifting or poor socket contacts
are the initial suspects so STOP if even
one write will not verify.

Mods

A look at the pinouts of these common
chips reveals how similar they all are. For
example, even if we can’t program a 2732
type, it is still possible to READ its con-
tents. The contents of the 2732 can then
be put in half of a 2764. When installing
the ‘‘double 32” connect the highest ad-
dress line to either 0 or 5 volts depending
on which half you want to use. These

S REM THE Z% EFPROMER
By D.TURNER + N.DENNIS
10 REM THIS ASSUMES THAT THE
COLE TO SE PROGRAMMED
i IS ALREADY STORED IN 9
UARIASLE CALLED D$.
i 15 REM IT SHOULD TAKE S@ SEC
LONGER TO EXECUTE 10088
LOCPS WITH THE EPROMER
ACTIVE (SCREEN FLIOKER>?
- THEN WITHOUT.
20 REM TYPICAL TINE FOR 1880
LOOPS OF THE PROGRAM
i IS 79 SECONDS.
THIS TIME OOES NOT __
i INCLUDE LINE 120 WHICH
i WOULD STOP THE TEST.
t
25 REM LINE 30 IS A DUMMY _FŪR
i TESTING ONLY. DELETE IT
AFTER TIMING OR IT WiLL
| DESTROY THE DATA IN C8.
| 30 CIM C$(1000)
50 REM
l 90 LET DEST=8192
| 99 SLOW
| 100 FOR I=1 TO 1002
; 110 POKE DEST+I,CODE O$(I) ;
: 420 IF PEEK (DEŚT+I):>CODE Osit
) THEN GOTO 9993
150 NEXT I
200 STO
599 PRINT “VERIFICATIGN ERROF A

Fiz. 3 4 simple BASIC listing for programming and compatible EPROM.

PAGE - 9 -

devices are cheap enough to make this
method the easiest way to get around
some tricky circuit construction.

Gotcha

The most common problem with project
or kit failure is faulty soldering. Before
powering up, go over the entire board
with a magnifying glass. Tiny wisps of
solder, while barely visible to the eye, can
cause expensive damage. Going over the
board thoroughly is time well spent. If
your project will not work, examine each
‘section independently. Remove the other
ICs from the board, and using the ‘‘How
it Works” section, trace the circuit’s
operation. With only the LS138 installed,
PEEKing 10000 should return 255. Any
other-result means that the ROM echo is
not suppressed; therefore, re-check the
138. The trap and the timer can be checked
statically, making them much simpler to
debug.

The next most likely trouble source is
the ZIF socket. A good socket wil cost
half of the total cost of the project, and is
worth every cent. The cheapies do not
guarantee the good connection that must
be maintained.

When programming, bear several
things in mind. The 21 volts needed to pro-
gram an EPROM is fatal to all other TTL
circuits. If this voltage varies by more
than a half a volt either way, the write
may not take. It is important that the
chosen supply can deliver at least S0mA
with good regulation. An on-card 5 volt
regulator and a decoupling capacitor
across every IC also helps maintain the ac-
curacy needed for programming. These
capacitors must not be eliminated or you
will be plagued with unpredictable results.

PARTS LIST

74LS 138 block decoder
74LS 10 triple Nand
74LS 123 dual timer

Resistors, Capacitors
200K 10 turn mini-trim type pot.
22K trim pot.

1N4009 or approx. equivalent
diodes

ZX edge connect and expansion strip

1.C. sockets (4-5 with personality)

Veroboard, Ribbon cable

A GOOD ZIP (Zero Insertion Force)

Socket

21 Voh power supply 29 Volt and

21.5 Volt (AA) batteries will work just fine

if fresh.

D1, D2

Software is likely to be the biggest additional expense over the life-
of the personal computer. This is the reason Bill Gates is the richest °
man in America. Many computers come from the factory loaded with
software, in some cases dozens of programs with an advertised value of
$1000.00 or more. In truth, many of the programs are likely to be
demonstration, outdated or "light" versions that lack key. features, so
take the monetary claims lightly. Then add to the budget to cover the
cost of upgrades.

It is not unheard of for a family to spend the equivalent of the
computer ’s purchase price for software over the life of the machine.
High-end productivity packages can cost hundreds of dollars, and the
Cool Game of the Month always seems to cost $50.00 or more. All that
software has to be backed up. The downside of having a gigabyte hard
drive is the cheapest way to back it up is on megabyte diskettes. To
save time and sanity, not to mention huge stacks of diskettes, consider
getting a Zip drive (about $150.00) or one of its removeable cartridge
drive cousins.

and there is the Internet, which turns your cash into electrons. If
your computer does not come with a 28 .8-kilobyte-per-second modem, or
something faster, plan to spend about $150.00 or more to add one.
These days, a computer without a modem is seriously underequipped.
Once the modem is installed, plan to budget at least $20.00 a month for

Internet access.

And then there are the hidden costs of the telephone lines. If some
family members plan to take advantage of unlimited Internet access,
other family members may be frustrated when the phone line is
constantly busy. A second phone line will ease the disputes, but
again, the phone company’s fingers will go walking through your wallet.
The unkindest cost of all is the long-distance charges spent for
waiting on line to talk with technical support.

The list goes on. There will be spare cables to buy at $10.00 to
$15.00 each, at least one exotic adapter for a couple of dollars, a
few subscriptions, membership dues for a local computer club and the
bribes paid to the neighbor ’s high-schooll kid to debug your mysterious
Windows.INI files. In the final accounting, when you access how mush
you have spent, the extra money for the "Dummies" books will seem quite

appropriate.

By Peter H. Lewis

i ce

ATTENTION LIST Subscribers: When it is time to renew your membership,
(look at your mailing label), please make out your check to Harvey
Rait, LIST President or to Robert Malloy, Treasurer. PLEASE DO NOT MAKE
OUT YOUR CHECK to LIST. Our bank requires a large amount of money in a
savings account in order to cash checks. THANK YOU!

Harvey Rait Robert Malloy
5 Peri Lane, 412 Pacific Street,
Valley Stream, NY 11581 Massapequa Park, NY 11762

Due to rising postage costs outside of the United States, we must raise
our annual dues accordingly:

CANADA and MEXICO $17.50 US, and the rest of the world $24.00 US.

Bob Malloy, LIST Treasurer
. PARE =- 14 =